Returns probabilities instead of text, so an AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → can automate an action or route to human review based on a threshold. Billing is on input tokens only, including images.
Key takeaways · AI-distilled
Liquid AI's d1-ultra is a multimodalA model that works with more than text — reading images, audio, or video, and sometimes generating them too.Full definition → decision model: instead of generating text it returns structured answers with probabilities via Boolean, Choice or Score question types, for tasks like visual defect inspection or checking damage claims against photos.
Several questions about the same input can go in one request, and the calibrationHow well a model's confidence matches reality — a calibrated model saying "90% sure" is right about 90% of the time.Full definition → probabilities are meant to decide when to automate an action versus send it to human review.
On Vercel AI Gateway it is called as liquid/d1-ultra, has a 32,768-token context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition →, and is billed on input tokens only, including images, with no output-token charge and no Gateway markup.
Terms in this piece · Glossary
multimodal — A model that works with more than text — reading images, audio, or video, and sometimes generating them too.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
calibration — How well a model's confidence matches reality — a calibrated model saying "90% sure" is right about 90% of the time.
